Today, we're carrying the sad news of the passing of Romain "roms" Liévin. It was relayed by François "mmu_man" Revol.

In this TI graphing calculators community, Romain was especially an instrumental part of the Linux Programmers Group, which brought the community pieces of libre, portable software (usually working on Linux, BSDs, macOS and Windows) for communicating with and emulating the TI graphing calculators we've come to love: libticonv, libtifiles(2), libticables(2), libticalcs(2), and their better-known TILP (II), TIEmu and TilEm front-ends respectively. The ability to communicate with (almost) the entire lineup of TI graphing calculators through a single, unified API, in its second iteration, is unmatched.

In 2007, he visited TI EdTech in Dallas, picking up a variety of calculators and cables to aid in further development of the communication libraries. TI doesn't want to invest in porting their software to Linux and the BSDs, and does usually no longer provide documentation, but helps select persons with hardware donations. Even to this day, few members of the user community had such a level of interaction with persons on TI's payroll.

Professionally, Romain was teaching some aspects of computer science and physics, and had additionally undertaken a PhD.

I met him several times over the years, and picked up his set of calculators and cables in 2010, in order to carry forward the maintenance and evolution of the LPG stack, mostly working on the libraries. A nice and smart person, with a good long-term vision for the community. It showed in decisions such as requiring that some features remain optional in TIEmu; this enabled TIEmu to remain usable into the 2020s with minimal maintenance, so despite the slight additional complexity it brought, time has now shown that it was the right way forward.

mmu_man will keep the memories from him at the engineering school, where we were young and technology was cool and interesting. That's when he remembers him smiling.

(Cross-post of https://www.ticalc.org/archives/news/articles/14/149/149491.html & https://ti-pla.net/t26959 )

Romain stopped working on the LPG stack in 2009. It's therefore natural that in 2025, even a number of community old-timers have never seen him post anywhere. Afterwards, he was still springing up infrequently on yAronet or TI-Planet.
In 2009, when he retired from the LPG stack, he was a father of definitely one but possibly two; three in the end. He was 45 years old.

Thinking back at it, I'm pretty sure that he was the first member of the TI graphing calculators community I met in person. I thought that was in 2005 or 2006 at the latest, but the e-mails I've just dug up indicate that it was in August 2004. I remember showing him how slow the graphics stack of Windows 9x/ME was, impacting TIEmu usage.
Not that more than 20 years later, I've met many members of the community. And AFAICR, Romain is the only one who came to my house, and the only one whose house I visited and met relatives. I met François more often than any other though - hard to beat working in the same room (part-time, he also worked at home, unlike me) for nearly a year.

It happens that a current member of the Casio community, "Lephe", is working as a PhD student in the same research lab as Romain was, though not in the same research team. He didn't know Romain well, and didn't know that Romain had a strong past in the calculators community, like most of their colleagues; on my side, I didn't know that Lephe was a PhD student there, either.
Small world, once again... that reminds me of François (then PhD student), I (MSc) and two colleagues (permanent researchers, back then and still nowadays) in that same research lab meeting Paul Bertrand, i.e. Adriweb's father, during a work trip at his (then) company in 2010 as well, pretty much in the time frame of me picking up 15 calculators from Romain.

The funeral service was attended by ~200 persons, I'd say. >> 100 but < 300.

I did not know this, this is sad.

Very sad news and thanks for reporting. RIP.

That's a name I recognize immediately from the link guide. A true giant in the community, with a lasting legacy. I hope his family can find peace, and can find some solace in knowing what a positive impact he's had on a pretty big group of people.

rest in peace

While Roman was on his way out of the community as I joined I likely would not be where I am now if not for his work on libti* and associated tools.

His legacy will live on in all the people his work has touched and helped even indirectly 16 years later and for many more to come.

I didn't know how many of the tools I use are owed to this man. Best wishes to his family and friends.

I'm really sorry to hear it. Like TheStorm, I am particularly aware of his work on linking (especially the Link Protocol Guide), but I'm sure he has touched my life in the calculator community in many other ways. May his memory be a comfort to his friends and family.

Captured in April of 2009, this page then said to be last updated in July of 2008 already mentioned two kids at the time: https://web.archive.org/web/20090403182852/http://www.lievin.net/index.html



Yes the world can be very small sometime. Well in fact it always is, it's just that most of the time we ignore it. Most of the time we live we people we already interacted with without knowing it, or who are just immediate relatives of them. This is why going to some IRL events is useful to help discover that (and it also helps to prevent the build-up of online resentment when there are disagreement, when you can actually relate them to real persons), like conferences, demo-parties…



I'm convinced Romain is one of the reasons why the whole TI ecosystem is lasting that much compared to some other calculator generations. Sustainability requires people to still continue the work, like Lionel, but this required people to build solid foundations that can last for a lot of time, like Romain did.

I wrote an article in French on LinuxFr.org that emphases on “He has touched my life in many other ways”:

• https://linuxfr.org/users/illwieckz/journaux/deces-de-romain-lievin-auteur-de-tilp-et-acteur-incontournable-de-la-scene-ti-sous-linux

One good example about unnoticed closeness is that someone mentioned in a comment that Romain also had an account on LinuxFr.org, and in fact he actually published there two articles. I had forgot about that myself…

I hadn't thought about it that way. That's not unreasonable, indeed.


I've never looked at the first iteration of the libti* APIs, but the second iteration makes it possible to target 3 protocol families (would be 4 if I had implemented the CX II's native NNSE, a wrapping for a variant of the "NSP" protocol officially dubbed NavNet, after working a bit on reverse-engineering NNSE) spanning the TI-Z80, TI-eZ80, TI-68k and Nspire series, communicating through basically all known cables, without the abstractions leaking too much outside the libti* APIs.
That's far from being bad work. Nearly 16 years after he passed the baton (mid-2009), the general architecture and most of the design remain almost unchanged, for good reason

In the beginning, the code base's layers seemed overly complex to me, but I fairly quickly realized that all layers are there for a reason. Remove any of these layers, and make some more or less common use case suffer. There are other, more lightweight implementations of TI's protocols; they're smaller because they're narrower.
If anything, an upper layer is missing for the highest level operations, as I realized later. I started creating that libtiopers over a decade ago, but just including a refactored version of TILP's tilp_calcs.cc makes the library basically too thin to stand in its own right...

Amusingly, even though I worked on the implementation of almost all areas of the libti* code, including tests, I somewhat suck at using the libti* API, especially libtifiles. It's the main technical reason why multiple large commits remain on the development branch of the tilibs repo: I want to include more functionality into the initial version of tifileutil.

Since there doesn't seem to be a way to attach files to posts on Cemetech, I'll link the document I uploaded at https://ti-pla.net/p277546 :
[img]https://tiplanet.org/forum/download/file.php?id=6499&mode=view[/img]

The message translates to something along the lines of
"
Death of Romain Liévin on the 2025/02/10

To Lionel Debroux and the whole T.I. community

A heartfelt thanks for your support in this ordeal we are enduring.
Romain was passionate about T.I. calculators and free software, especially Linux.
You were alongside him throughout these years of binge thinking and coding !
Let's hope that his work keeps inspiring you for what follows.
Very sincerely,
<signed by Romain's 3 children and their mother>
"

Dang. : (